Hacker News Reader

The thing is, an open ecosystem that allows users to run any software they want on their devices already exists, including on Apple devices, in the form of the Web.

And despite all of its flaws it's the most successful software deployment platform the world has ever seen, to the point where even Apple can't afford to not include it on their platforms when we all know they'd rather not, because it's simply too valuable to consumers and would make their devices obsolete if they didn't include it.

There's no central gatekeeper for the web, and sure, it has its dark corners, but it has not devolved into absolute chaos as you so adamantly suggest all such ecosystems would, and continues to deliver such an incredible amount value to consumers and businesses alike.

The web was only possible because it was developed in an era before everybody was using closed down devices where the manufacturer dictates what software you can run on it. And at this rate, if the web ever falls out of favor (and Apple for one is doing everything it can to make sure it does), you can definitely be sure that nothing like it will ever be allowed to exist again.

As a community of founders and makers, I'm sure we can all imagine what a sad world it would be to live in if we had to first convince some platform gatekeeper that our idea is worthwhile and make sure our ideas don't conflict with their interests before being allowed to turn our ideas into reality and deliver them to users, let alone having to pay a cut of our revenues to them for the privilege.

This is probably the best comment on the subject I have ever read, for whatever that's worth.

The web doesn't run on its own platform. It does run on browsers, which are controlled by few careful instances. Mobile browsers are subject to the same rules as any other app and most popular browsers are owned by lower level platforms, separated from website/webapp owners. It is not the same as side-loading. Web analogy would work if websites were executable binaries or if browsers were much less restrictive. It is an open ecosystem under a strictest environment ever made. Nobody is going to download your random binary as mindlessly as they tap on a-hrefs.

Following the idea you present, Epic should just start making browser games with in-game purchases just over a credit card. Why doesn't it then? What's wrong with the web that Epic couldn't just publish on playfortnite.com and that'd be it?

By running your products via a browser, you're both handing more control over to Apple via Safari, the platform they'd have to run on (only Apple can create a web browser on iOS remember), and leaving power/performance/features on the table for your competitors (ex: netflix/youtube can only stream lower resolution content via the web vs via their apps).

Thanks for your comment. I want to clarify some things.

> There's no central gatekeeper for the web, and sure, it has its dark corners, but it has not devolved into absolute chaos as you so adamantly suggest all such ecosystems would, and continues to deliver such an incredible amount value to consumers and businesses alike.

I never stated that all open systems would turn into chaos. What I suggested is that systems with unchecked advanced AI capabilities probably will. Technology is changing and old paradigms are not guaranteed to work in the future.

You seem to be looking only at the present and there I find your sentiment somewhat reasonable. However, looking at the future, safety will become much more important. If the software you are deploying has the potential to (unintentionally) kill people, shut down cities, or otherwise wrack havoc because you haven’t completely mastered the art of training machine intelligence in a safe way... you simply don’t want this to be openly sold, same as you don’t want to buy medicine or medical technology from a 5 year old kid you just met on the subway. You also don’t have the right to hurt people to fulfill your desire to make things.

I know what I am referring to requires some looking down the road but we will get there in due time. More regulation is (hopefully) inevitable.

What I still think is open for discussion are the exact terms of such vetting/control. It might very well be that terms will get less onerous. That there will be requirements for gate keepers to implement. However, freedom from control is not going to be the long term solution and this is not a bad thing!

We are just talking about handheld personal computers. I don‘t know why people always try to come up with car comparisons when its really about computers, but: You already can modify your car today, including the (engine) software, but have to make sure it is still street legal. I sure can imagine certifications for autonomous cars and their software. So if you can and want to tinker with it, be prepared to pay for the certification to keep it street legal.

I won‘t argue about the rest of your comment about AI, as I think this leads to discussing castles in the air.

I guess the point is that if we are talking about laws it seems that there is a desire to abstract to broader classes of things that are easily identifiable. I was pretty broad and generalized to technology running programmable software.

If you want to stay with personal computing equipment (whatever that may mean exactly) the point still holds if you look a little bit down the line. You want someone responsible for harm that is created by harmful software and as a user you want an easy way to get save software. Both is much easier with an entity curating and signing vetted software.

What is interesting in this case is the question who would be responsible for this and what are the rules that would need to be followed?

I think it is difficult to untangle the vetting position from manufacturing because the manufacturer is likely the best expert on the computing platform. But other arrangements could be tried out. In terms of rules to follow I am sympathetic to some general rules that vetting should follow but much thought would need to be put into what those rules should be. My prime concern would be safety and minimization of great harm. But how does one go about this? Even external payment providers could be argued to be a potential source of harm if they are not vetted and certified. It‘s tricky!

I think a possible solution would be to make costs and revenue for app stores transparent and stipulate that margins should remain reasonable. Maybe set up a developer board that has a voice in the app store policy making similar to how employee representatives have a seat on the board in many German companies.

I think the line is “does this harm a human who does not consent.” Regulating autonomous vehicles? Apple being able to reject apps that steal user data? Both within reason.

But rejecting simply because it lets someone pay in an alternate manner crosses a line.

While I'm inclined to agree with the spirit of your argument, I believe Apple makes the argument that their payment gateway enforcement is in fact in line with that “does this harm a human who does not consent” test.

The argument goes, if I'm not mistaken, that by enforcing their payment gateway, they're assuring that users aren't handing over credit card info and other PPI in an insecure manner.

Whether you buy that or not is up to you, but this is definitely a defence I've heard.

> The argument goes, if I'm not mistaken, that by enforcing their payment gateway, they're assuring that users aren't handing over credit card info and other PPI in an insecure manner.

Doesn't Apple specifically have a deal with Amazon to ignore the 30% cut for them?

I believe they only consider ‘digital goods’ to be in-app purchases. For instance you can’t buy kindle books or streaming video content through the amazon app(s).

> Would you also argue that you have the right to run any software on it?

Duh, yes. It's like asking if i can drive my (human-driven) car anywhere. Just because it's "technology" doesnt mean all human agency goes out the window.

It seems like you are not thinking things through then. You would at least need to get some kind of certification that your car remains „street legal“. You cannot seriously expect to run any autonomous driving software that you want in your car?

obviously the software should have some guarantees, just like cars have safety standards. But those should not be too strict to stifle competition, nor should companies be allowed to say "it s illegal to modify your car's firmware"

And how would you guarantee that the combination of parts that you are combining are safe? This would mean that there would need to be standards throughout the car.

If all of the construction would be modularized with open interfaces, I could imagine something like this working... So, I must admit that it seems theoretically possible to set something like this up but we are currently nowhere near this. Every car is a black box that only the manufacturer really knows. Same for software. What you demand would require all software modules to be “standardized” (think API with safety requirements and guarantees) so that automatic verification could take place.

But what are appropriate “module” sizes? Do we regulate every function call? Just applications? What if Apple sells the iphone plus software as one application? What are generic rules you could use to decide what is the “right” application to regulate? By what mechanisms can we come to good decisions around this?

It’s an interesting vision... but also totally different from the world we live in today. It’s not as simple as you make it sound.

Apple already has a similar practice in place for desktop apps with Notarization, does it not?

I don’t see how we can’t get to a point where we have something similar for iOS. They can limit the APIs they have access to and the User has the choice.

The question is if all computing platform should be forced to offer this possibility or not. I don‘t think it‘s clear cut and some kind of „vetted“ side loading seems like one possible solution. You would still need to expect to pay Apple for the vetting, though.